[Low-bit optim] Add Llama2-7B finetune benchmarks#746

msaroufim merged 5 commits into

pytorch:mainpytorch/ao:mainfrom

gau-nernst:update_optim_benchgau-nernst/ao:update_optim_benchCopy head branch name to clipboard